Monte Carlo project estimation · Python · alpha

Stop guessing dates.
Plan with distributions.

Planaco models every task as a probability distribution and runs thousands of Monte Carlo simulations, so you ship honest confidence ranges instead of a single optimistic number.

$ pip install planaco
View on GitHub
The problem

A single number hides the risk

“This takes 5 days” throws away everything you know about uncertainty. A distribution keeps that knowledge and tells you how often you'll actually be late.

~ 21 days
One bar. No sense of best case, worst case, or how confident you should be.
21 days is a coin flip, 26 is near-certain
The full shape of outcomes, with percentile markers you can actually commit to.
Live demo

Try it. No install needed.

Four tasks, real dependencies. Every drag runs a fresh 10,000-trial Monte Carlo, exactly what planaco does.

How long until project completion?

10,000 simulations · days to completion
A coin flip—daysP50
Safe to promise—daysP85
Near-certain—daysP95
 the same model, in planaco
project = Project(name="Web App", unit="days")
project.add_task(design)                              # 5 · 7 · 14
project.add_task(frontend, depends_on=[design])       # 8 · 12 · 15
project.add_task(backend,  depends_on=[design])       # 10 · 15 · 25
project.add_task(testing,  depends_on=[frontend, backend])
project.statistics(n=10_000)   # → P50, P85, P95
How it works

Three steps from tasks to confidence

1

Model each task

Give a min / most-likely / max, or pick one of six distributions. Planaco captures the uncertainty you already feel.

2

Link dependencies

Declare what blocks what. Planaco builds a DAG and computes the critical path on every simulated run.

3

Simulate & decide

Run 10,000 Monte Carlo trials in milliseconds. Read off percentiles, confidence intervals, and the riskiest tasks.

From YAML to forecast

Define it once, simulate forever

from planaco import Task, Project

project = Project(name="Web App", unit="days")

design   = Task(name="Design",   min_duration=5,  mode_duration=7,  max_duration=14)
frontend = Task(name="Frontend", min_duration=8,  mode_duration=12, max_duration=15)
backend  = Task(name="Backend",  min_duration=10, mode_duration=15, max_duration=25)
testing  = Task(name="Testing",  min_duration=2,  mode_duration=3,  max_duration=5)

project.add_task(design)
project.add_task(frontend, depends_on=[design])
project.add_task(backend,  depends_on=[design])
project.add_task(testing,  depends_on=[frontend, backend])

stats = project.statistics(n=10_000)
print(stats["percentiles"]["p85"])
Six distributions

Pick the shape that fits the uncertainty

Ship a range you can stand behind.

$ pip install planaco
GitHub PyPI